Cannabis use disorder (CUD), which is up to ten times more common in patients with schizophrenia (SCZ) than in the general population, worsens the course of this severe psychiatric disorder. Since SCZ occurs in 1% of the population, the co-occurrence of CUD in 13% to 42% of people with this disorder presents society with an important public health problem. Unfortunately, most antipsychotics available for treatment of patients with SCZ do not appear to limit their cannabis use. Moreover, the one antipsychotic that preliminary data suggest may well limit cannabis use in these patients, clozapine (CLOZ), is not used for this purpose; it is reserved for patients whose psychosis is treatment resistant. The overarching idea behind this proposal, however, is that CLOZ's use is being unreasonably restricted and should be made more widely available for patients with SCZ who have a co-occurring CUD but whose psychosis is not necessarily treatment resistant. This notion is supported by our preliminary clinical and animal data on the effects of CLOZ, as well as our neurobiological model of the basis of cannabis use in patients with SCZ that provides a pharmacologic rationale for this effect of CLOZ. Even given all the arguments favoring the potential benefits of CLOZ in patients with SCZ and CUD, however, its side effect profile will likely limit its use until a fully powered study demonstrates its ability to decrease cannabis use n patients with SCZ. This proposal aims to launch such a study. If, as we hypothesize, this study confirms and extends our previous preliminary data of the effects of CLOZ in patients with SCZ and CUD, it will provide a strong impetus to expand the use of CLOZ in this population. In the proposed study, 132 patients who are comorbid for both SCZ and CUD will be randomized to a 12 week treatment course with either CLOZ or risperidone (RISP). The primary specific aim of this proposal is: (1) To test the hypothesis that patients treated with CLOZ will have decreased cannabis use as compared to patients treated with RISP. Subsidiary aims will further elucidate the effects of CLOZ in this population: (2) a) To determine whether patients treated with CLOZ will have improvements in (i) psychiatric symptoms; (ii) quality of life; and (iii) neuropsychological functions as compared to those taking RISP; and b) to explore whether patients taking CLOZ will show improved reward responsiveness as compared to those taking RISP; and (3) To explore whether those patients with the val/val genotype at the COMT Val158Met locus are more likely to decrease cannabis use during CLOZ treatment than are those without the val/val COMT genotype. Should this study indicate that CLOZ will lessen cannabis use in patients with SCZ more than RISP, it will provide evidence needed to begin to shift clinical practice toward its use in these patients. Given the increased morbidity associated with CUD in patients with SCZ, doing so could dramatically improve the clinical outcome of these individuals. Lastly, CLOZ's use in this study may also reflect its potential to serve as a prototype of the next generation of medications for treatment of SCZ and co-occurring CUD.